基于WSN的旋转机械设备故障时频监测方法  

Time Frequency Monitoring Method of Rotating Machinery Equipment Fault Based on WSN

作　　者：孙留存 胡从川 钱大龙 SUN Liucun;HU Congchuan;QIAN Dalong(China Green Development Investment Group Co.,Ltd.,Beijing 100020,China)

机构地区：[1]中国绿发投资集团有限公司,北京100020

出　　处：《机械与电子》2024年第3期76-80,共5页Machinery & Electronics

摘　　要：由于旋转机械设备结构和振源较为复杂,以单一故障经验设置的阈值无法准确分解多模态故障,为提升故障监测效果,提出基于WSN的旋转机械设备故障时频监测方法。引入集合经验模态分解故障时频信号,分解不同时刻的振动信号,计算IMF分量的能量,结合归一化能量指标和IMF矩阵奇异谱熵指标,完成旋转机械设备故障时频信号分解。根据特征分解结果,运用训练后免疫RBF神经网络监测旋转机械设备故障。实验结果表明,该方法能够缩短监测时间、提高故障监测准确率。Due to the complex structure and vibration source of rotating machinery equipment,the threshold set by single fault experience cannot accurately decompose multi-modal faults.In order to improve the fault monitoring effect,a time-frequency monitoring method for rotating machinery equipment faults based on WSN is proposed.The time-frequency signal of fault is decomposed by the collective empirical mode,the vibration signal at different times is decomposed,the energy of the IMF component is calculated,and the normalized energy index and the IMF matrix singular spectrum entropy index are combined to complete the decomposition of the fault time-frequency signal of rotating machinery equipment.According to the results of feature decomposition,the trained immune RBF neural network is used to monitor the faults of rotating machinery.The experimental results show that this method can shorten the monitoring time and improve the fault monitoring accuracy.

关 键 词：集合经验模态 旋转机械设备 故障监测 时频监测 主成分分析 RBF神经网络 

分 类 号：TH17[机械工程—机械制造及自动化]